DeepSeek-V4-Flash-W4A16-FP8-MTP

canada-quant/DeepSeek-V4-Flash-W4A16-FP8-MTP

Text Generation50.7B parametersvllmLicense: mit
Tracked by OpenModelStats since September 7, 2026View on Hugging Face ↗
Downloads 30D
202
Total downloads
34.7K
Gained 7D
+35(increase)
0.1%(up) vs baseline
Gained 30D
—
Likes
20
+0 in 7D
Rank · tracked models
#39,071
605(down 605 places)this week
Parameters
50.7B
Last source update
May 30, 2026
Momentum
24.2
  • DeepSeek-V4-Flash-W4A16-FP8-MTP gained 35 downloads during the last seven tracked days.
  • It moved from #11383 to #11527 among tracked Text Generation models this week.

Downloads over time

Cumulative total downloads observed by OpenModelStats

Likes over time

Rank among tracked models

Lower is better

Overview

DeepSeek-V4-Flash-W4A16-FP8-MTP is a text generation model published by canada-quant. OpenModelStats has tracked the model since Sep 7, 2026, most recently observing it Oct 6, 2026.

Published
May 22, 2026
Last updated
May 30, 2026
Library
vllm
License
mit
Task
Text Generation
Parameters
50,709,699,698
Spaces
0
Derivative models
—
Velocity 7D/day
5
vllmsafetensorsdeepseek_v4deepseekcompressed-tensorsw4a16gptqfp8mtpspeculative-decodingmixture-of-expertsmoe

Model family

Based on base-model relationships reported by the source's metadata.

Similar models

canada-quant/DeepSeek-V4-Flash-W4A16-FP8-MTP — Downloads, Growth & Stats | OpenModelStats